Project AI Rules Setup

Overview

This skill creates a structured AI rule system for a project:

• AGENTS.md at project root: project overview, goals, file structure, common workflows, key resources.
• .codebuddy/rules/*.md: granular, enforceable rules for AI behavior (identity, data sources, compliance, images, workflow).

Rules are extracted from ~/.workbuddy/MEMORY.md (user-level preferences) and the current workspace context, then written into files so any AI assistant opening the project can behave consistently.

When to Use

• User asks to "write project rules", "create AGENTS.md", "set up .codebuddy/rules", "把规则落到文件里".
• A project has accumulated implicit conventions that should be made explicit for AI collaborators.
• Starting a new project and want to bootstrap AI behavior rules from user memory.

Workflow

Step 1: Read User Memory

Read ~/.workbuddy/MEMORY.md to extract:

• Identity rules (how to address user, AI persona)
• Data source rules (what data is allowed/forbidden)
• Compliance rules (advertising law, medical compliance)
• Image generation rules (supplier, model, watermark removal)
• Material priority rules (IMA, local asset packages)
• Workflow rules (when to use which skill, mandatory checks)

If ~/.workbuddy/MEMORY.md does not exist, skip and rely on conversation context.

Step 2: Inspect Project Structure

List the project root directory to understand:

• Project name and purpose
• Existing files and conventions
• Existing skills/scripts used
• Output directories

Step 3: Create .codebuddy/rules/ Directory

Create the directory if it does not exist:

mkdir -p .codebuddy/rules